Drug Information Journal - March 2009 - (Page 112) 112 SPECIAL FEATURE Foote tion about your topic. Translation: it is not necessary to tell the readers of Drug Information Journal about the structure and duties of the Food and Drug Administration, unless the article is “Structure and Roles of the Food and Drug Administration.” The introduction should engage the reader and must provide the research question. Even a review article or opinion article has a research question. Many research articles benefit from the addition of a rationale for the hypothesis formulated from the original research question (eg, “Because drug X has been shown to be efficacious in the setting of disease Y, we were interested to know if similar efficacy would be seen in the setting of disease Z”). A good introduction sets the stage and does so in three to four paragraphs (3). METHODS The only articles that might not have a methods section are opinion articles. If your article is a critical review of a topic, your method of finding and selecting published literature must eliminate bias. Some guidance on this aspect has been published (4,5). The reader cannot judge the validity of your work if adequate methods are not provided. A caveat: the methods section seems straightforward, but care should be taken to avoid a cookbook approach, for example, “First, do this; then add that. Then do this.” In general, the methods section benefits greatly from subheadings, which will come in handy in the results section (6). RESULTS The results section should provide just the results, without commentary or excuses, and should present them in the same order as the methods (7). Keep the reader in mind: as an author, you have seen the data repeatedly; as a naive, first-time reader, I have not. One of the secondary endpoints may well be the most exciting finding of your research, but you must report your results—all of your results—in a logical order, logical being the order presented in the methods section. You may not omit results you do not like, but you may say “data not shown” for results that are not statistically significant and not the objective of the study. I have a bias against the use of the term trend, which I find to mean that the data are not statistically significant and that the authors are using subjective analyses other than the prespecified analysis to try to salvage their work. The data are either statistically significant or they are not, which is why we include P values and confidence intervals. DISCUSSION The discussion section should be a balanced interpretation of the data. Although the discussion section should summarize the data, it should not repeat it. You should compare and contrast your findings with the work of others and you must return to the research question posed in the introduction and show how your research supported—or did not support—the question. All studies have limitations, which may not be solely inherent in the study reported but may be due to the choice of study design, and these should be discussed. If you are reporting original research, do not overstate or overextend your results beyond the population studied. Using the example of drug X again, the results of your present study may be good in the setting of disease Y and disease Z, but have not been studied in disease W, so no statement of usefulness can be made in the setting of disease W. However, you now have your rationale for your hypothesis and research question for your next study: “Because drug X has been shown to be efficacious in the settings of disease Y and disease Z, we were interested to know if similar efficacy would be seen in the setting of disease W.” The discussion should not introduce new ideas or data and generally should not suggest future studies. In more than one article, a small company has enthusiastically provided their clinical plans, only to have a larger, better staffed company use their ideas to engage investigators and patients. The author, however, may wish to use the results of the study being reported to comment on the implications of the study findings. It must be noted that the article is about
For optimal viewing of this digital publication, please enable JavaScript and then refresh the page. If you would like to try to load the digital publication without using Flash Player detection, please click here.